U.S. patent number 8,272,710 [Application Number 12/174,061] was granted by the patent office on 2012-09-25 for bi-directional print masking.
This patent grant is currently assigned to Eastman Kodak Company. Invention is credited to Douglas W. Couwenhoven, Richard C. Reem, Christopher Rueby, Kevin E. Spaulding.
United States Patent |
8,272,710 |
Spaulding , et al. |
September 25, 2012 |
Bi-directional print masking
Abstract
A method for reducing banding artifacts for bi-directional
multi-pass printing on an inkjet printer utilizing a printhead with
a plurality of ink nozzles includes defining different print masks
to be used for leftward and rightward printing passes such that
both the order of ink laydown and the timing between ink laydown on
different passes are each substantially constant for a given
horizontal position within the image, independent of the vertical
position within the image; and printing an input image on the
inkjet printer with the defined print masks using a bi-directional
multi-pass print mode.
Inventors: |
Spaulding; Kevin E.
(Spencerport, NY), Couwenhoven; Douglas W. (Fairport,
NY), Reem; Richard C. (Hilton, NY), Rueby;
Christopher (North Chili, NY) |
Assignee: |
Eastman Kodak Company
(Rochester, NY)
|
Family
ID: |
41529962 |
Appl.
No.: |
12/174,061 |
Filed: |
July 16, 2008 |
Prior Publication Data
|
|
|
|
Document
Identifier |
Publication Date |
|
US 20100013878 A1 |
Jan 21, 2010 |
|
Current U.S.
Class: |
347/15 |
Current CPC
Class: |
G06K
15/107 (20130101); B41J 19/142 (20130101) |
Current International
Class: |
B41J
2/205 (20060101) |
Field of
Search: |
;347/14,15 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Peng; Charlie
Attorney, Agent or Firm: Spaulding; Kevin E.
Claims
The invention claimed is:
1. A method for reducing banding artifacts for bi-directional
multi-pass printing on an inkjet printer utilizing a printhead with
a plurality of ink nozzles comprising: a) defining different print
masks to be used for leftward and rightward printing passes such
that the order of ink laydown is constant for all vertical and
horizontal positions within the image, and wherein the timing
between ink laydown on different leftward and rightward passes is
constant for all vertical positions within the image at a
particular horizontal position; and b) printing an input image on
the inkjet printer with the defined print masks using a
bi-directional multi-pass print mode; wherein if the print mode has
an even number of printing passes, the print mask defined for
leftward passes uses a first set of active ink nozzles that
excludes a first plurality of contiguous ink nozzles on one end of
the printhead, the first plurality of contiguous ink nozzles
extending to and including the ink nozzle at the extreme end of the
printhead, and the print mask defined for rightward passes uses a
second set of active ink nozzles that excludes a second plurality
of contiguous ink nozzles on the other end of the printhead, the
second plurality of contiguous ink nozzles extending to and
including the ink nozzle at the extreme other end of the printhead;
and wherein if the print mode has an odd number of printing passes,
the print mask defined for one printing direction uses a first
larger set of printing nozzles and the print mask defined for the
other printing direction uses a uses a second smaller set of
printing nozzles together with two subsets of non-printing ink
nozzles, one on either end of the printhead, the subsets of
non-printing ink nozzles extending to and including the ink nozzles
at the respective extreme ends of the printhead.
2. The method of claim 1 wherein if the print mode has an even
number of printing passes, the number of unused ink nozzles in the
first plurality of contiguous ink nozzles and the number of unused
ink nozzles in the second plurality of contiguous ink nozzles are
substantially equal to the number of active ink nozzles in the
first and second sets of active ink nozzles divided by the number
of printing passes.
3. The method of claim 1 wherein the inkjet printer is a color
inkjet printer having at least two different ink colors.
4. The method of claim 1 further including defining a first page
advance distance to be applied before leftward printing passes and
a second different page advance distance to be applied before
rightward printing passes.
5. The method of claim 1 wherein the inkjet printer is a
multi-level inkjet printer having a plurality of printing levels,
and the different print masks to be used for leftward and rightward
printing passes are defined by specifying a plurality of mask
planes, each of which is associated with a different printing
level.
6. The method of claim 5 wherein the mask planes are selected by
multitone code values from a multitoning operation.
7. The method of claim 1 wherein the different print masks to be
used for the leftward and rightward printing passes are defined by
rearranging segments of a conventional multi-pass print mask.
8. The method of claim 1 wherein at least one of the different
print masks to be used for the leftward and rightward printing
passes is defined to have a non-uniform duty cycle across the
height of the print mask.
9. The method of claim 8 wherein page advance distances are defined
to produce an overlap printing region where additional printing
passes are used to reduce boundary artifacts.
Description
FIELD OF THE INVENTION
This invention pertains to the field of inkjet printing systems,
and more particularly to a method for reducing banding artifacts
associated with bi-directional multi-pass printing on an inkjet
printer.
BACKGROUND OF THE INVENTION
A typical inkjet printer reproduces an image by ejecting small
drops of ink from a printhead containing ink nozzles, where the ink
drops land on a receiver medium (typically paper) to form ink dots.
Inkjet printers typically reproduce color images by using a set of
color inks, usually cyan, magenta, yellow, and black. It is well
known in the field of inkjet printing that if ink drops placed at
neighboring locations on the page are printed at the same time,
then the ink drops tend to flow together on the surface of the page
before they soak into the page. This can give the reproduced image
an undesirable grainy or noisy appearance often referred to as
"coalescence". It is known that the amount of coalescence present
in the printed image is related to the amount of time that elapses
between printing adjacent dots. As the time delay between printing
adjacent dots increases, the amount of coalescence decreases,
thereby improving the image quality. There are many techniques
present in the prior art that describe methods of increasing the
time delay between printing adjacent dots using techniques referred
to as "interlacing", "print masking", or "multi-pass printing".
These methods often involve advancing the paper by an increment
less than the printhead width for each printing pass. As a result,
successive passes or "swaths" of the printhead overlap, which has
the additional advantage that it can help to reduce one-dimensional
periodic artifacts referred to as "bands" or "banding" that can
result due to clogged or misdirected ink nozzles. See, for example,
U.S. Pat. Nos. 4,967,203 and 5,992,962. The term "print masking"
generically means printing subsets of the image pixels in multiple
partially overlapping passes of the printhead relative to a
receiver medium.
Another attribute of modern inkjet printers is that they typically
possess the ability to vary (over some range) the amount of each
ink that is deposited at a given location on the page. Inkjet
printers with this capability are referred to as "multitone" inkjet
printers because they can produce multiple density tones at each
location on the page. Some multitone inkjet printers achieve this
by varying the volume of the ink drop produced by the nozzle by
changing the electrical signals sent to the nozzle or by varying
the diameter of the nozzle. See for example U.S. Pat. No.
4,746,935. Other multitone inkjet printers produce a variable
number of smaller, fixed size droplets that are ejected by the
nozzle, all of which are intended to merge together and land at the
same location on the page. See for example U.S. Pat. No. 5,416,612.
These techniques permits the printer to vary the size or optical
density of a given ink dot, which produces a range of density
levels at each location, thereby improving the image quality.
Another common way for a multitone inkjet printer to achieve
multiple density levels is to print a small amount of ink at a
given location on several different passes of the printhead over
that location. This results in the ability to produce a greater
number of density levels than the nozzle can fundamentally eject,
due to the build up of ink at the given location over several
passes. See, for example, U.S. Pat. No. 5,923,349.
In U.S. Pat. No. 5,790,150, Lidke et al. disclose a method where
multiple passes are made over the page while fractionally advancing
the page. In each pass, the pattern of dots in the data swath is
constructed with sufficient spacing between the dots such that the
printhead can be scanned across the page at a velocity that is
higher than the firing frequency limit of the nozzles.
In U.S. Pat. No. 6,310,640, Askeland discloses a print masking
method in which nozzles at the ends of the printhead print with
lower duty than nozzles near the center of the printhead, thereby
reducing the possibility of banding artifacts occurring at the
boundaries between successive printed swaths.
In U.S. Pat. No. 6,206,502 Kato et al. also discloses a method for
reducing the duty for nozzles at the ends of the printhead. This
method involves using a page advance which is smaller than the
number of nozzles in the printhead divided by the number of passes,
so that there is a region at the ends of the printhead where the
passes overlap for an additional pass. The goal of this is to hide
artifacts that can result at the boundaries of the printhead due to
page advance errors, etc. Vinals et al disclose a similar method in
U.S. Pat. No. 6,375,307 for a single-pass printing configuration.
These print masking methods are sometimes referred to as
"fractional print masking" in the literature.
In U.S. Pat. No. 6,238,037, Overall et al. disclose a print masking
method for a multilevel inkjet printer in which the print mask
contains a set of threshold values. A dot will print at a given
location on a given pass if the multitone code value for that pixel
is greater than the threshold for that pass. This method requires
that if a dot gets printed at a given pixel on pass N, then it also
must receive dots on passes 0 through N-1.
In U.S. Pat. No. 6,454,389, Couwenhoven et al. disclose a print
masking method suitable for multilevel inkjet printers that can
produce multiple sized ink drops.
In U.S. Patent Application Publication No. 2007/0201054, which is
incorporated herein by reference, Billow et al. disclose a print
masking that utilizes a print mask having a plurality of mask
planes, each mask plane corresponding to a multitone code value.
This approach has the advantage that dot patterns printed in
response to different multitone levels can be independent from each
other.
The method of Billow et al will now be described in more detail to
illustrate print masking. Turning to FIG. 1, a typical inkjet
printer system is shown in which an image preprocessor 20 receives
a digital image from a host computer 10, and performs standard
image processing functions such as sharpening, resizing, color
conversion, and multitoning to produce a multitoned image signal i.
The multitoned image signal i is composed of a set of color data
planes hereinafter referred to as color channels. Each color
channel corresponds to a particular colorant in the printer, such
as the cyan, magenta, yellow, or black inks used in a typical
inkjet printer. The data including each color channel is a two
dimensional array (width=w, height=h) of individual picture
elements, or "pixels". The pixel's location in the image is
specified by its (x,y) coordinates in the array, where
0.ltoreq.x.ltoreq.w-1 and 0.ltoreq.y.ltoreq.h-1. The x location of
the pixel is also referred to as the pixel column number, and the y
location of the pixel is referred to as the pixel row number. The
term "signal" is used to generically refer to the array of pixels
having digital code values that form the image.
A swath data generator 30 then receives the multitoned image signal
i and generates a swath data signal s, which controls the volume of
ink printed by an inkjet printhead (or printheads) 40. The process
of print masking is contained within the swath data generator 30.
Prior to multitoning, each pixel contains a numeric code value
(typically on the range {0,255}) for each color channel that
indicates the amount of the corresponding colorant to be placed at
the given pixel's location in the image. After multitoning (at the
output of the image preprocessor 20), the image is represented by
multitone code values, where the range of pixel code values has
been reduced to match the number of density levels that the inkjet
printer can produce. For binary inkjet printers, the possible
multitone code values will be either 0 or 1, indicating whether to
print 0 or 1 drops of ink. Multitone inkjet printers will accept
multitone code values on the range {0,N-1}, where N is the number
of possible multitone code values, and is normally the number of
density levels (or number of drops) that the multitone inkjet
printer can produce at a given pixel.
Turning now to FIG. 2, the details of the swath data generator 30
are shown. A "swath" of data is defined as the dot ejection data
that is required during one motion of the printhead across the
page. In FIG. 2, according to the method of Billow et al., a print
mask for a given color contains a set of mask planes 50, 52, 54,
56, each of which has a M.sub.w.times.M.sub.h array of individual
mask elements 60.
Often, the mask height M.sub.h will be equal to the number of
nozzles in the printhead, although this is not a fundamental
restriction, and a mask height of lesser or greater value can be
used. One of the mask planes is selected for a given pixel
according to the multitone code value of the multitoned image
signal i, as shown in FIG. 2. A pixel column index x.sub.m and a
pixel row index y.sub.m are computed according to the following
equations: x.sub.m=x % M.sub.w (1) y.sub.m=y % M.sub.h (2) where x
is the pixel column number and y is the pixel row number of the
current pixel being processed, M.sub.w is the mask width, M.sub.h
is the mask height, and the "%" symbol indicates the mathematical
modulo operator. The value of the swath data signal s is then
determined by selecting a mask element 62 from the chosen mask
plane according to: s=MaskPlane(i,x.sub.m,y.sub.m) (3)
Turning now to FIG. 3, an example mask plane 70 is shown. In the
mask plane 70, each of the individual mask elements 80 can be one
of two values: a first value (0) indicating that no ink drop is to
be ejected, and a second value (1) indicating one drop of ink is to
be ejected. Thus, if the mask plane 70 corresponds to multitone
code value 1, and a uniform 8.times.8 input image of multitone code
value 1 was input to the swath data generator 20, then a dot
pattern indicated by the mask elements having value "1" in the mask
plane 70 would be printed in one pass of the printhead. For
purposes of illustration, the mask plane 70 is shown as having a
mask width and mask height of 8, although one skilled in the art
will recognize that a mask of any arbitrary size can be used.
Generally, mask sizes will be significantly larger than this, with
the mask height typically being equal to the height of the
printhead.
Turning now to FIG. 4, the dot patterns resulting from three
subsequent passes of an inkjet printhead having 8 nozzles in
response to a uniform 8.times.8 input image of multitone code value
1 are shown. In this example, the print mask used has the mask
plane 70 of FIG. 3 set to correspond to multitone code value 1, and
the receiver media is advanced by four raster lines between each
pass of the printhead. Since the input image has a uniform field of
multitone code value 1, mask plane 70 will be selected for every
pixel in the 8.times.8 image, and the pattern of dots printed in
each of the three successive swaths will correspond to the pattern
of 1s in the mask plane 70. The resulting swath patterns 90, 92, 94
are shown offset horizontally from each other, and the resulting
pattern of ink dots 96 is shown, produced by overlapping the
individual swath patterns. Note that in regions where two
successive print passes overlap, every pixel location has received
one drop of ink, which corresponds to the desired output for the
8.times.8 input image of multitone code value 1. Thus, the print
mask shown in the example is appropriate for use in a "2-pass"
printmode, meaning that two passes of the printhead are required
for the desired final dot patterns to be printed. This also means
that the mask plane 70 is designed such that the top half and
bottom half of the mask, when overprinted on two subsequent print
swaths, will produce the desired number of ink drops at each pixel.
In this case, this implies that the top half and the bottom half of
the mask plane 70 are complementary, such that a single ink drop
will be printed at each location.
Often inkjet printers are configured to print in a bi-directional
print mode, where ink is applied as the printhead moves in both
rightward and leftward directions. A common problem with inkjet
printers that utilize bi-directional multi-pass printing is that
they can be susceptible to banding artifacts caused by differences
in the order of ink laydown and the timing between ink laydown on
different passes. These differences can cause systematic variations
in the produced color due to interactions between the ink and
media. For example, consider FIG. 5, which illustrates a
bi-directional 2-pass print mode. A printhead having an associated
print mask 100 first moves from left-to-right, printing ink in a
first swath 101 (indicated by a pattern of upward sloping diagonal
lines). The paper is then advanced by half of the printhead height
and the printhead makes a second pass over the paper, this time
moving from right-to-left in a second swath 102 (indicated by a
pattern of downward sloping diagonal lines). The paper is then
advanced again, and the printhead prints a third swath 103
(indicated by a pattern of vertical lines), moving from
left-to-right. The print mask 100 is labeled with two sections, 1
and 2. Section 1 corresponds to the part of the printhead that
prints on the media the first time it passes over a given region of
the page. Likewise, section 2 corresponds to the part of the
printhead that prints on the media the second time it passes over a
given region of the page.
It can be seen that there are differences in timing between ink
laydown on different print passes, both across the page, as well as
down the page. For example, consider a first overlap region 104
where ink is first applied during the rightward first swath 101,
and then during the leftward second swath 102. In a left portion of
the first overlap region 105 there will be a relatively long time
delay between the times that ink is applied on the first swath 101
and the second swath 102. This is due to the fact that the
printhead must travel all the way across the page, then turn around
and come all the way back across the page. However, in a right
portion of the first overlap region 106 there will be a relatively
short time delay between the times that ink is applied on the first
swath 101 and the second swath 102. This is because the printhead
needs to travel a shorter distance before it turns around and comes
back. The reverse is found to be true in a second overlap region
107 where ink is first applied during the leftward second swath
102, and then during the rightward third swath 103. In a left
portion of the second overlap region 108 there will be a relatively
short time delay between the swaths, whereas in a right portion of
the second overlap region 109 there will be a relatively long time
delay between the passes.
The differences in the time delays, both across the page and from
swath-to-swath, can result in significant differences in the
characteristics of the reproduced image. When there is a longer
time delay between passes, the ink applied during the first pass
will have a longer time to dry or soak in to the paper. This can
result in noticeable differences in the density of the printed
region. Additionally, there can also be noticeable differences in
the image structure characteristics. For example, coalescence
artifacts, as well as surface characteristics such as gloss and
haze, are often observed to be a function of the timing between
when neighboring ink drops are applied. Typically, the
swath-to-swath differences at a given horizontal position in the
image are much more objectionable than the variations across the
page. This is because the swath-to-swath differences produce a
periodic artifact where the image characteristics vary for
alternating swaths. These artifacts are sometimes referred to as
bi-directional banding artifacts since they are inherently related
to bi-directional multi-pass print modes. The magnitude of these
artifacts is quite dependent on the characteristics of the
particular ink and media used in the inkjet printer, as well as
print mode attributes such the amount of ink, the number of passes
and the printing speed. The magnitude of the bi-directional banding
artifacts can even be affected by the size of the image since this
can have an effect on how long it takes the printhead to travel
across the page. For many combinations of ink, media and print
mode, the bi-directional banding artifacts have been found to be
quite objectionable.
Bi-directional banding artifacts can be even more severe for the
case of printing color images with multiple color inks. Consider
the case of a color inkjet printer using cyan, magenta, yellow and
black inks. Typically, each ink will be printed using a different
column of nozzles in the printhead. Therefore, as the printhead
moves back and forth across the paper, there will also be
differences in the order that the different color inks are applied.
For example, consider the case where a uniform blue image region is
to be printed using equal amounts of cyan and magenta inks. If the
cyan nozzles are located to the right of the magenta nozzles in the
printhead, they will be applied before the magenta drops on a
rightward swath, but after the magenta drops on a leftward
pass.
Consider the case where a blue image region is printed using the
2-pass configuration shown in FIG. 5. In the left portion of the
first overlap region 105 the ink will be applied as
cyan/magenta/long delay/magenta/cyan. However, in the right portion
of the first overlap region 106 the ink will be applied as
cyan/magenta/short delay/magenta/cyan. Similarly, in the left
portion of the second overlap region 108 the ink will be applied as
magenta/cyan/short delay/cyan/magenta, and in the right portion of
the second overlap region 109 the ink will be applied as
magenta/cyan/long delay/cyan/magenta. Therefore, at a horizontal
position on the left edge of the page, the image will alternate
back and forth between regions that vary in both the order of ink
laydown and the timing between ink laydown on different passes.
This can cause objectionable periodic bi-directional banding
artifacts that vary in both color (lightness, hue and/or chroma)
and image structure. Generally the magnitude of the bi-directional
banding will decrease towards the center of the page where the
contributions to the bi-directional banding that result from
differences in the timing between ink laydowns will be negligible,
leaving only the contributions that result from differences in the
order of ink laydown. The magnitude of the bidirectional banding
will then increase again for horizontal positions toward the right
side of the page, although the phase of the banding artifacts that
result from differences in the timing between ink laydowns will be
reversed relative to the left side of the page. The magnitude of
the bi-directional banding artifacts is often much more
objectionable for image regions printed with two or more colored
inks than it is for image regions using only a single ink since the
reproduced color can be strongly influenced by which ink drops are
printed on top.
A variety of methods have been proposed to alleviate the
objectionable bidirectional banding artifacts. One solution is to
use only uni-directional print modes where ink is only printed when
the printhead is moving in one direction (e.g., rightward).
However, this solution significantly limits the throughput of the
printer since it is necessary to wait for the printhead to return
back to the starting position before printing another swath.
Another step that can be taken is to slow down the printing speed
to give the ink/media interactions more time to stabilize between
passes, but this too will significantly impact the printer
throughput. Note that while this approach can help reduce the
contributions of the bi-directional banding that are due to
differences in the timing between ink laydowns, it will not
alleviate the contributions that result from differences in ink
ordering.
Another way to reduce the magnitude of the bidirectional banding
artifacts is to increase the number of printing passes. This will
effectively slow down the rate of ink deposition and dilute the
impact of the differences in ink order and timing at any given
location. However, this will also have a direct impact on the
throughput of the printer, so it is not a desirable solution for
applications where print speed is a critical requirement.
Another solution that has been proposed is to modify the amount of
ink laydown for the rightward and leftward printing passes. One way
to accomplish this is to use different color transforms to process
the image for the different printing passes. For example, see U.S.
Pat. Nos. 6,354,692 and 7,054,034, and U.S. Patent Application
Publication No. 2003/0048327. Alternatively, an ink depletion
operation can be used to modify the amount of ink that is printed
depending on the print direction. One way that this can be
accomplished is to modify the print masks as described in U.S. Pat.
No. 6,545,773. These methods can help alleviate the component of
the bi-directional banding that results from differences in ink
ordering. However, they will be ineffective at compensating for the
contributions of the bidirectional banding that result from
differences in the timing between ink laydowns since these effects
will vary from left to right across the page, and these methods do
not provide for changing the ink laydown as a function of the
horizontal position.
SUMMARY OF THE INVENTION
The present invention represents a method for reducing banding
artifacts for bidirectional multi-pass printing on an inkjet
printer utilizing a printhead with a plurality of ink nozzles
comprising defining different print masks to be used for leftward
and rightward printing passes such that both the order of ink
laydown and the timing between ink laydown on different passes are
each substantially constant for a given horizontal position within
the image, independent of the vertical position within the image;
and printing an input image on the inkjet printer with the defined
print masks using a bi-directional multi-pass print mode.
It is an advantage of the present invention that images with
reduced bidirectional banding artifacts can be produced for inkjet
printers using bi-directional multi-pass print modes.
It has the additional advantage that it will reduce bi-directional
banding artifacts that result from both differences in the order of
ink laydown, as well as the timing between ink laydown on different
passes.
It is yet another advantage of the present invention that the
reduction in bidirectional banding artifacts can be achieved
without a significant impact on the printing speed. In some
embodiments of the present invention, the impact on the printing
speed will be negligible.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a flow diagram showing an typical inkjet printer
system;
FIG. 2 is a diagram illustrating print masking according to the
present invention;
FIG. 3 is a diagram showing the details of a mask plane;
FIG. 4 is a diagram illustrating multi-pass printing;
FIG. 5 is a diagram illustrating overlapping print swaths;
FIG. 6 is a diagram illustrating a 2-pass print masking
configuration in accordance with the present invention;
FIG. 7 is a diagram illustrating a pair of mask planes;
FIG. 8 is a diagram illustrating a 6-pass print masking
configuration in accordance with the present invention;
FIG. 9 is a diagram illustrating a 5-pass print masking
configuration in accordance with the present invention;
FIG. 10 is a diagram illustrating a 2-pass print masking
configuration using different page advance distances in accordance
with the present invention; and
FIG. 11 is a diagram illustrating a 2-pass print masking
configuration using fractional masking in accordance with the
present invention;
DETAILED DESCRIPTION OF THE INVENTION
The present invention represents a method to reduce bi-directional
banding artifacts typically associated with bi-directional
multi-pass printing on an inkjet printer by way of a novel method
for defining different print masks for leftward and rightward
printing passes. The leftward and rightward print masks are defined
such that the order of ink laydown and the timing between ink
laydown on different passes are substantially constant for a given
horizontal position within the image, independent of the vertical
position within the image.
Turning now to FIG. 6, the method of the present invention will be
described. A page 119 is printed on an inkjet printer using a
bi-directional multi-pass print mode by moving a printhead
horizontally across the page in a first direction (e.g. rightward),
firing ink drops from nozzles on the printhead as it passes over a
receiver media such as paper. The paper is then advanced by some
distance less than the height of the printhead, and the printhead
is then moved back across the head in the other direction firing
additional ink drops. This process is repeated until the entire
image is printed. This particular example illustrates a 2-pass
print mode, where, for any given region of the page, ink is printed
during two different printing passes. However, as will be described
later, this method can be generalized to print modes with any
number of printing passes.
The movement of the printhead across the page is typically referred
to as a printing pass, or sometimes as a print swath. FIG. 6
illustrates a sequence of four printing passes: a first printing
pass 120 where the printhead is moved across the page in a
rightward direction; a second printing pass 121 where the printhead
is moved across the page in a leftward direction; a third printing
pass 122 where the printhead is again moved across the page in a
rightward direction; and a forth printing pass 123 where the
printhead is again moved across the page in a leftward
direction.
An important feature of the present invention is that different
print masks are used for the rightward and leftward printing
passes. A rightward print mask 124 is used for the rightward
printing passes (i.e., the first printing pass 120 and the third
printing pass 122). A leftward print mask 125 is used for the
leftward printing passes (i.e., the second printing pass 121 and
the fourth printing pass 123). The rightward print mask 124 and the
leftward print mask 125 are designed in a manner such that the
order of ink laydown and the timing between ink laydown on
different passes are substantially constant for a given horizontal
position within the image independent of the vertical position
within the image.
In one preferred embodiment of the present invention, the rightward
print mask 124 and the leftward print mask 125 are defined by
starting from a conventional multi-pass print mask and rearranging
segments of the print mask to form two new print masks. The example
shown in FIG. 6 illustrates how the print masks can be formed from
conventional 2-pass print masks, such as the mask plane shown in
FIG. 3. The rightward print mask 124 includes 3 segments
corresponding to different subsets of the printing nozzles, one
segment labeled as "X" and the other two segments labeled as "1".
The segment labeled as "X" corresponds to a set of unused ink
nozzles on one end of the printhead (in this case the top end). The
segments labeled with a "1" are formed by extracting a segment of a
conventional 2-pass print mask and inserting it into the rightward
print mask 124 in two different positions. The segment of the
conventional 2-pass print mask that is extracted corresponds to the
part of the printhead that prints on the media the first time it
passes over a given region of the page. For the print mask shown in
FIG. 3, this would correspond to the lower half of the print mask
(pixel row indices 4-7). The leftward print mask 125 also includes
3 segments corresponding to different subsets of the printing
nozzles, one segment labeled as "X" and the other two segments
labeled as "2". The segment labeled as "X" again corresponds to a
set of unused ink nozzles, this time on the other end of the
printhead. The segments labeled with a "2" are formed by extraction
a segment of a conventional 2-pass print mask and inserting it into
the leftward print mask 125 in two different positions. The segment
of the conventional 2-pass print mask that is extracted corresponds
to the part of the printhead that prints on the media the second
time it passes over a given region of the page. For the print mask
shown in FIG. 3, this would correspond to the upper half of the
print mask (pixel row indices 0-3).
To understand how the method of the present invention satisfies the
required condition that the order of ink laydown and the timing
between ink laydown on different passes are substantially constant
for a given horizontal position within the image, independent of
the vertical position within the image, consider a first overlap
region 126 and a second overlap region 127. The first overlap
region 126 corresponds to the region of the page where the first
printing pass 120, the second printing pass 121 and the third
printing pass 122 all overlap. The second overlap region 127
corresponds to the region of the page where the second printing
pass 121, the third printing pass 122 and the fourth printing pass
123 all overlap.
Consider the scenario discussed above relative to FIG. 5 where it
is desired to print a blue patch using a printhead with a column of
cyan nozzles that are located to the right of a column of magenta
nozzles. In a left portion of the first overlap region 128 ink will
be applied in the first printing pass 120 and the second printing
pass 121 as cyan/magenta/long delay/magenta/cyan. In a left portion
of the second overlap region 129 ink will be applied in the third
printing pass 122 and the fourth printing pass 123, again as
cyan/magenta/long delay/magenta/cyan. Similarly, in a right portion
of the first overlap region 130 ink will be applied in the first
printing pass 120 and the second printing pass 121 as
cyan/magenta/short delay/magenta/cyan. And in a right portion of
the second overlap region 131 ink will be applied in the third
printing pass 122 and the fourth printing pass 123, again as
cyan/magenta/short delay/magenta/cyan.
It can be seen that for the left portion of the first overlap
region 128 and the left portion of the second overlap region 129,
ink is applied such that the order of ink laydown and the timing
between ink laydown on different passes is the same
(cyan/magenta/long delay/magenta/cyan) for a horizontal position on
the left side of the page, independent of the vertical position
within the image. Likewise, it can also be seen that for the right
portion of the first overlap region 130 and the right portion of
the second overlap region 131, ink is applied such that the order
of ink laydown and the timing between ink laydown on different
passes is also the same (cyan/magenta/short delay/magenta/cyan) for
a horizontal position on the right side of the page, independent of
the vertical position within the image. Furthermore, it can be seen
that the same would be true for all other horizontal positions
across the image as well. Therefore, the abrupt changes in color
and/or image structure than can occur for a conventional multi-pass
print masking configuration due to the pass-to-pass differences in
the order of ink laydown and the timing between ink laydown on
different passes are eliminated. This drastically reduces the
bi-directional banding artifacts that are often observed for
bi-directional multi-pass printing.
It should be noted that the timing between ink laydown on different
passes is different for the right side and the left side of the
page, and therefore the color and image structure characteristics
can be somewhat different from side-to-side across the image.
However, any changes in the image characteristics will occur slowly
across the width of the image, with no abrupt transitions. As a
result, these differences will generally be much less objectionable
than the bi-directional banding artifacts associated with
conventional multi-pass print masking configurations, such as that
illustrated in FIG. 5. As a result, the image quality is
significantly improved relative to bi-directional multi-pass
printing using conventional multi-pass print masks.
A more detailed illustration of how a rightward print mask 124 and
a leftward print mask 125 can be defined starting from a
conventional multi-pass print mask is illustrated in FIG. 7. For
this example, the 2-pass mask plane shown in FIG. 3 is used as the
seed pattern for forming the new print masks. A rightward mask
plane 132 includes 3 segments corresponding to different subsets of
the printing nozzles: two identical subsets of active nozzles 134
at the bottom and middle of the rightward mask plane 132, and a
subset of unused nozzles 135 at the top of the rightward mask plane
132. The subsets of active nozzles 134 correspond to the segments
of the rightward print mask 124 labeled as "1" and the subset of
unused nozzles 135 corresponds to the segment of the rightward
print mask 124 labeled as "X" in FIG. 6. The subsets of active
nozzles 134 are formed by extracting the bottom four rows (labeled
with pixel row indices y.sub.n=4-7) of the conventional multi-pass
mask plane 70 shown in FIG. 3. The four rows of the rightward mask
plane 132 corresponding to the subset of unused nozzles 135 are
filled with zeros so that no drops will be printed by these rows of
nozzles.
Similarly, a leftward mask plane 133 includes 3 segments
corresponding to different subsets of the printing nozzles: two
identical subsets of active nozzles 136 at the top and middle of
the leftward mask plane 133, and a subset of unused nozzles 137 at
the bottom of the leftward mask plane 133. The subsets of active
nozzles 136 correspond to the segments of the leftward print mask
125 labeled as "2" and the subset of unused nozzles 137 corresponds
to the segment of the leftward print mask 125 labeled as "X" in
FIG. 6. The subsets of active nozzles 136 are formed by extracting
the top four rows (labeled with pixel row indices y.sub.n=0-3) of
the conventional multi-pass mask plane 70 shown in FIG. 3. The four
rows of the leftward mask plane 133 corresponding to the subset of
unused nozzles 137 are filled with zeros so that no drops will be
printed by these rows of nozzles. These print masks are designed to
be used with a page advance distance 138 of four pixels. It can be
seen that if the rightward mask plane 132 and the leftward mask
plane 133 are overlaid after applying the four pixel page advance,
every pixel will have a "1" because of the fact that the top and
bottom halves of the conventional multi-pass print mask plane 70
were designed to be complementary.
While the process of defining the rightward print mask 124 and the
leftward print mask 125 based on a conventional multi-pass print
mask represents a simple method for implementing the present
invention, it will be obvious to one skilled in the art that there
are many different methods that could be used to form the rightward
print mask 124 and the leftward print mask 125 to achieve the
desired result that the order of ink laydown and the timing between
ink laydown on different passes are substantially constant for a
given horizontal position.
In a preferred embodiment of the present invention, rightward print
mask 124 and the leftward print mask 125 are defined by specifying
a plurality of mask planes, each associated with a different
printing level, as was taught in U.S. Patent Application
Publication No. 2007/0201054. However, the print masks can also be
defined using any of a number of other techniques for specifying
where drops should be printed on each printing pass, such as by
specifying threshold matrices as was taught in U.S. Pat. No.
6,238,037.
The example described with reference to FIG. 6 corresponds to the
case of a 2-pass print mode. The method of the present invention
can easily be generalized to higher numbers of printing passes as
well. For example, a 6-pass print mode configuration is illustrated
in FIG. 8, which uses a rightward print mask 140, and a leftward
print mask 141. In this case, the print masks are formed by
breaking a conventional 6-pass print mask into six segments labeled
by numbers 1 through 6. Print mask segment 1 corresponds to the
portion of the print mask that controls printing the first time the
printhead passes over a given region of the media (i.e., the bottom
1/6 of the print mask). Print mask segment 2 corresponds to the
portion of the print mask that controls printing the second time
the printhead passes over a given region of the media (i.e., the
next 1/6 of the print mask up from the bottom), and so on. The
rightward print mask 140 is formed, starting from the bottom, by
using two of each of the odd-numbered print mask segments (1, 3 and
5) followed by a segment labeled as "X" corresponds to a set of
unused ink nozzles on the top end of the printhead. Similarly, the
leftward print mask 141 is formed, starting from the bottom with a
segment labeled as "X" corresponds to a set of unused ink nozzles,
followed by two of each of the even-numbered print mask segments
(2, 4 and 6). Like the 2-pass print mode example given in FIG. 6,
it can be seen that the use of the rightward print mask 140 and a
leftward print mask 141 for a 6-pass bi-directional print mode
ensures that the order of ink laydown and the timing between ink
laydown on different passes are substantially constant for a given
horizontal position within the image independent of the vertical
position within the image.
FIG. 9 shows a configuration with a rightward print mask 150 and a
leftward print mask 151 that can be used to implement a 5-pass
print mode according to the method of the present invention. As can
be seen, the formation of the print masks is somewhat different for
print modes with an odd number of print passes. The print mask
defined for one printing direction (rightward in this case) uses a
larger set of printing nozzles than the print mask for the other
printing direction. The print mask having the smaller set of
printing nozzles includes two subsets of non-printing nozzles, one
on either end of the printhead. In this example, the rightward
print mask 150 is formed, starting from the bottom, by using two of
each of the odd-numbered print mask segments (1, 3 and 5), but does
not contain any segments of unused ink nozzles. The leftward print
mask 151 is formed, starting from the bottom with a segment labeled
as "X" corresponding to a set of unused ink nozzles, followed by
two of each of the even-numbered print mask segments (2 and 4),
followed by a second segment labeled as "X" corresponding to a set
of unused ink nozzles on the top end of the printhead.
While the configurations described above for implementing the
present invention accomplish the desired result of substantially
reducing objectionable bi-directional banding artifacts, they do
have an impact on the throughput of the inkjet printer. This is
largely due to the fact that there are a significant number of
unused ink nozzles in the leftward and rightward print masks. For
the case of an even number of printing passes, it can be seen that
the number of unused nozzles is substantially equal to number of
active nozzles divided by the number of passes. For example, from
FIG. 6 it can be seen that for a 2-pass print mode, the number of
unused ink nozzles will be approximately 1/2 the number of active
ink nozzles. Likewise, from FIG. 8 it can be seen that for a 6-pass
print mode, the number of unused ink nozzles will be approximately
1/6 the number of active ink nozzles. For the case of an odd number
of printing passes, the print mask for one printing direction will
use all of the printing nozzles, but for the other direction there
will be two subsets of unused ink nozzles. If the two printing
directions are averaged, the number of unused nozzles can be seen
to still be substantially equal to number of active nozzles divided
by the number of passes.
The result of having these subsets of unused ink nozzles is that
the page advance distance is correspondingly smaller for a given
printhead length than it would be using conventional multi-pass
print masks with print modes having the same number of printing
passes. Essentially, the configurations that have been described
are equivalent to adding an extra printing pass. Therefore, the
throughput for a 2-pass print mode of the embodiment described
above would be approximately equivalent to the throughput for a
conventional 3-pass print mode. This would represent a decrease in
the throughput to 2/3 (67%) of what it would be for the
conventional 2-pass print mode. As the number of passes increases,
the throughput impact is less significant. It can be seen that the
throughput reduction factor will be approximately equal to N/(N+1),
where N is the number of printing passes. For example, for a 6-pass
print mode using the embodiment described above, the throughput
would only be 6/7 (86%) of a conventional 6-pass print mode. In
many cases, the image quality advantages of the present invention
will be well worth the reduction in throughput.
For print modes with an even number of printing passes, the
throughput impact can be largely mitigated by using an alternate
embodiment of the present invention that uses different page
advance distances before the leftward and rightward printing
passes. This is illustrated in FIG. 10, which represents a
variation of the 2-pass print mode that was shown in FIG. 6. This
approach takes advantage of the fact that the number of active ink
nozzles in the rightward print mask 124 and the leftward print mask
125 are the same. Therefore, there is no reason to devote 1/3 of
the printhead to unused ink nozzles. In the example shown in FIG.
10, the print masks use a substantially smaller number of unused
ink nozzles. In this configuration, the printhead make a first
printing pass 160 in the rightward direction. The page is then
advanced a first page advance distance 166 corresponding to the
number of unused ink nozzles at the top of the rightward printing
mask 164 before making a second printing pass 161 in the leftward
direction using a leftward print mask 165. The page is then
advanced by a second page advance distance 167 corresponding to the
number of active ink nozzles in the rightward printing mask 164 and
the leftward printing mask 165. This process is then repeated for
the third printing pass 162 and the fourth printing pass 163, and
so on.
In the limit, the number of unused nozzles can be shrunk to zero.
In this case, the first page advance distance 166 would be zero,
and the second page advance distance 167 would be the full width of
the printhead. This would produce a throughput equivalent to the
conventional 2-pass print masking configuration. However, it would
have the disadvantage that it eliminates any redundancy in the
printing nozzles since the same nozzle would pass over a given
position on the paper in both printing directions. As a result,
there would be no way to compensate for artifacts associated with
missing or misdirected nozzles, which would eliminate one of the
advantages of multi-pass printing. Therefore, it will generally be
desirable to use a first page advance distance 166 of at least a
few pixels. If the number of unused nozzles is maintained at a
small value, there will be a negligible impact on throughput
relative to the conventional multi-pass print masking
configuration, while maintaining all of the advantages of the
present invention with respect to reducing the visibility of
bi-directional banding artifacts. It should be noted that this
approach for improving the throughput will only work for the case
of an even number of printing passes, so there will be an advantage
to using print modes having an even number of printing passes when
implementing the present invention.
It will be obvious to one skilled in the art that the present
invention can be combined with many different print masking
methods. For example, in U.S. Patent Application Publication No.
2007/0201054, Billow et al. describe using print masks having a
non-uniform duty cycle, meaning that not all of the nozzles in the
printhead will print with the same duty (see FIG. 7 and FIG. 8 of
U.S. Patent Application Publication No. 2007/0201054). Generally,
the duty of the nozzles near the ends of the printhead will be
lower than the duty in the center of the printhead, and the masks
will be defined in such a way that the sum of the duties for the
print masks over all passes will be a constant, independent of the
row of the image. This approach can be advantageous for hiding
banding artifacts that commonly occur near the swath boundaries.
Such masks can easily be rearranged and used to form leftward and
rightward print masks according to the methods described earlier.
However, this can cause abrupt changes in the duty cycle across the
height of the print mask, which cannot always be desirable.
Alternately, it is possible to generate leftward and rightward
print masks having a non-uniform duty cycle that satisfy the
requirement of the present of invention that the order of ink
laydown and the timing between ink laydown on different passes are
substantially constant for a given horizontal position within the
image. However, this requires that the leftward and rightward print
masks be defined directly rather than simply rearranging
conventional multi-pass print masks.
The method of the present invention can also be combined with
fractional print masking techniques, such as that disclosed by U.S.
Pat. No. 6,206,502. Consider the arrangement shown in FIG. 11,
which represents a variation of the 2-pass print mode with unequal
page advance distances that was shown in FIG. 10. In this
configuration, the printhead makes a first printing pass 170 in the
rightward direction. The page is then advanced a first page advance
distance 176 corresponding to the number of unused ink nozzles at
the top of the rightward printing mask 174 before making a second
printing pass 171 in the leftward direction using a leftward print
mask 175. The page is then advanced by a second page advance
distance 177. This process is then repeated for the third printing
pass 172 and the fourth printing pass 173, and so on. In this case,
the second page advance distance 177 is less than the number of
active ink nozzles in the rightward printing mask 174 and the
leftward printing mask 175. As a result, this creates a 2-pass
printing region 178 where the image is printed with the expected 2
printing passes, as well as an overlap printing region 179 where
the edges of the print masks overlap to provide additional printing
passes.
The rightward printing mask 174 is labeled with several different
segments. A segment labeled as "X" corresponds to a set of unused
ink nozzles on the top end of the printhead. A segment labeled "1"
corresponds to the main printing region of the print mask which
prints ink the first time the printhead passes over a given region
of the paper. Fractional mask segments labeled "1F" and "3F"
correspond to the overlap printing region 179 where the image is
printed with more than 2 printing passes. The fractional mask
segment labeled "1F" will print using a reduced duty the first time
the printhead passes over a given region of the paper, and the
fractional mask segment labeled "3F" will print using a reduced
duty the third time the printhead passes over a given region of the
paper. Likewise, the leftward printing mask 175 is also labeled
with several different segments. A segment labeled as "X"
corresponds to a set of unused ink nozzles on the bottom end of the
printhead. A segment labeled "2" corresponds to the main printing
region of the print mask which prints ink the second time the
printhead passes over a given region of the paper. Fractional mask
segments labeled "2F" and "4F" correspond to the overlap printing
region 179 where the image is printed with more than 2 printing
passes. The fractional mask segment labeled "2F" will print using a
reduced duty the second time the printhead passes over a given
region of the paper, and the fractional mask segment labeled "3F"
will print using a reduced duty the fourth time the printhead
passes over a given region of the paper. Generally, the duty of the
fractional mask segments will be defined in such a way that the sum
of the duties over all the printing passes will be a constant,
independent of the row of the image.
The advantage that the fractional print mask configuration shown in
FIG. 11 has over other configurations is that it will generally be
less sensitive to boundary artifacts that are apt to occur at the
edges of the printhead due to sources such as variations in the
page advance distance. However, it can be seen that the order and
timing of the ink laydown in the overlap printing region 179 will
be different than in the 2-pass printing region 178. As a result,
this can produce some variations in the color and/or image
structure of the printed image, mitigating some of the improvements
in the bi-directional banding artifacts. The objectionability of
the boundary artifacts needs to be traded off against the
objectionability of the bi-directional banding artifacts for a
given printing configuration to determine the optimal tradeoff.
Often the visibility of the boundary artifacts can be reduced
substantially using an overlap printing region 179 that is only a
few printing nozzles wide without producing objectionable
bi-directional banding artifacts.
The invention has been described in detail with particular
reference to certain preferred embodiments thereof, but it will be
understood that variations and modifications can be effected within
the spirit and scope of the invention.
Parts List
10 host computer 20 image preprocessor 30 swath data generator 40
inkjet printhead 50 mask plane 52 mask plane 54 mask plane 56 mask
plane 60 mask element 62 mask element 70 mask plane 80 mask element
90 swath pattern 92 swath pattern 94 swath pattern 96 ink dots 100
print mask 101 first swath 102 second swath 103 third swath 104
first overlap region 105 left portion of the first overlap region
106 right portion of the first overlap region 107 second overlap
region 108 left portion of the second overlap region 109 right
portion of the second overlap region 119 page 120 first printing
pass 121 second printing pass 122 third printing pass Parts List
Cont'd 123 fourth printing pass 124 rightward print mask 125
leftward print mask 126 first overlap region 127 second overlap
region 128 left portion of the first overlap region 129 left
portion of the second overlap region 130 right portion of the first
overlap region 131 right portion of the second overlap region 132
rightward mask plane 133 leftward mask plane 134 subsets of active
nozzles 135 subset of unused nozzles 136 subsets of active nozzles
137 subset of unused nozzles 138 page advance distance 140
rightward print mask 141 leftward print mask 150 rightward print
mask 151 leftward print mask 160 first printing pass 161 second
printing pass 162 third printing pass 163 fourth printing pass 164
rightward printing mask 165 leftward printing mask 166 first page
advance distance 167 second page advance distance Parts List Cont'd
170 first printing pass 171 second printing pass 172 third printing
pass 173 fourth printing pass 174 rightward printing mask 175
leftward printing mask 176 first page advance distance 177 second
page advance distance 178 2-pass printing region 179 overlap
printing region
* * * * *